130 The Plant World. 



as to secondary hybridizations "svith either parent species or 

 Avith various recombinations. 



The chief fact of interest in the present connection was 

 that of the possible reaction af the various types to cHmate. To 

 test this a series of tAvelve young trees representing the total 

 range of leaf-variation %vas taken to the Desert Laboratory at 

 Tucson, Arizona, and exposed to the extremes of the arid sub- 

 tropical climate at that place. In addition to the direct action 

 of the meteoric factors, the soil was of course widely different 

 from that of the habitat of the plant on Staten Island, since 

 soils are so largely a function of climate. 



First of all it was noted that these introduced oaks were 

 attacked l)v the desert rodents which found their lea.es much 



ViH 1. — Plciiitlct-^ from progeny of tree of Qui reus heterophylla. J hears leaves not dij 

 tinguishable from thoje of Q. Phellos, and 17 resembles Q. rubra- //. UI, J^', anrt V form 

 a series between / and 1'/ — Pholognpheci Anril, 1906. 



more to their taste than those of the oaks natixe on theadjacenl 

 mountain slopes. This fact, however, would not have operated 

 verv stronglv as a selective factor. The hot, dry fore-summers 

 and arid after-summers were the real selective agencies deter- 

 mining survivals, and the notes on the cultures illustrate the 

 fate of such a ])rogeny very clearly. 



The entire lot of plants survived tlie seasons of 1907 and 

 had cast their leaves with the frosts of December, being noted 

 as entirely dormant on Dec. 10. In 1*^)08 the individuals near 

 the middle of the series rej)resenting the trees usually passing 

 under the name of Q. hchrophy/la were latest in awakening. 



Elimination began in 1^.09 and by June \?>, only five plants 

 of the narrower leaved types survived. On Oct, 22, but four 



